We need to be careful with how GenAI is used: PhonePe CTO Rahul Chari

We’ve been comfortable saying that in a fully connected network where there is no barrier to entry… if you take the number of programs that exist and are launched, there is a decision that has been made over time and in the background. about consumer decisions based on reliable information.

Starting at the end of 2019 we stopped the refunds and incentive programs for users, it is the consumer’s choice in such a place. For the companies that have invested time, effort and money along with the network to create this great digital payment system, I believe that the hat is not heard.

Any type of cup is offensive to any part. The question is why future investment, innovation and business will happen at all. Placing a cap is not a good example of a systemic vulnerability. A model with stricter standards based on scale would be one option. We invest heavily in technical development… to ensure that performance, reliability and availability are always at the forefront.

But the directors are not infallible either. Market stability is dangerous

The regulator’s concerns are fair. The question is whether the answer is correct. These problems can be solved in many ways. The fact that there are new players constantly coming in and becoming important (major) players shows that there is an appetite to enter and there is an opportunity to build new use cases. Having something like a cup can be frustrating – it can go against the main reason.

The issue of merchant discount (MDR) is being discussed recently but nothing has been done on the part of the regulators.

We have been very open about this. Zero MDR on UPI played a role in promoting UPI, especially on the legal side. I think that at some point the salary for the salary should be self-sustaining and should be encouraged to build a company that pays the salary. I believe that the MDR should be done at some point so that the unit’s finances and wages are good.

How is PhonePe using GenAI in its processes?

When it comes to GenAI we need to be careful with its use, especially when it comes to controlled environments. We’ve been using ML (machine learning) extensively in risk and fraud detection for a very long time. At our rate of over 270 million daily, I can’t use the established rules to even begin to determine if the job should go through or not. We’ve been using ML a lot for cluster determination, for climbing.

GenAI – we need to be careful because we need to be able to explain when you are doing things like underwriting. Even in coding, I think one should be more careful when it comes to money management or when it comes to using financial facilities.

How is PhonePe’s technology prepared to manage scale?

A big part of our efforts on the payment side continues to be growth and performance. More than 50-70 percent of the work goes into the continuous improvement of our systems. It also comes down to having the best team, having the best talent on large distributed systems is essential. One of the things that made us happy – this is our third startup – we built a network of amazing talent in our first go and with Flipkart. Some of our engineers and architects continue to be with us. We made a plan that the first day will be giving 10 million dollars. That was our goal on the first day. There are many decisions we made in the past that have helped us. One of the other decisions we made was to stay on-prem, so the payments are not running on the public cloud. This gives us the ability to use more advanced tools and make many of our systems work better outside of legal requirements such as data interpretation. We use the public cloud for some purposes.

Fraud and cyber threats on UPI are on the rise. How is PhonePe planning for this?

Many of our funds are vulnerable and areas of fraud detection. This is one of our platforms that we have developed as Guardian and now we are looking to see if we can put it out into the environment with other players. Whether it’s doing device fingerprinting to ensure there’s no device manufacturing, or having a platform where we can run our ML models and implement rules at scale… matching to see if the work being done matches the user experience. This may be through behavioral technology or any other malicious attempt, we try to prevent this.

There are several approaches to cybersecurity. We have two different functions one group looks from outside and are like ethical hackers and another infosec group looks inside and changes regularly etc.

The more UPI has grown, the more we have to deal with it (cyber threats). This is important because the next 300 million users are the ones that need the most abuse and anti-spoof education.

There is an argument for using software made in India more than MNCs. What are your thoughts?

We have built our systems entirely on open source software. We have been able to use some of the best work that people have done in the world in building population reduction systems at a very low cost without being tied to any solutions. We have just opened our three solutions. PhonePe Github is now a public repository. All of our orchestration solutions are developed in-house and open source. And we open up our security system again…and hope to give more back to the community. I believe that since we have used so much from open source we should give it back to the community.

First Published: Oct 13 2024 | 11:36 PM IS